An erroneous coupling of a cable to a relay device such as a switch, a failure in a relay device, or the like may cause a loop in a data link layer in a network. When a loop occurs, the flooding of broadcast frames increases a load on the network, and may lead to a state in which communication is difficult. Accordingly, various methods that detect and suppress a loop have been disclosed.
For example, a cable modem system has been proposed in which a center device generates a loop detecting frame that detects a loop, and determines that a cable modem has detected a loop when the loop detecting frame is supplied from a subscriber side (for example Japanese Laid-open Patent Publication No. 2000-183943 and the like). Further, a loop detecting device has been disclosed which determines that a loop has occurred in a transfer path in a network when there is a receiving port where an amount of traffic of broadcast frames or multicast frames exceeds a threshold value (for example Japanese Laid-open Patent Publication No. 2012-175184 and the like).
Even when the occurrence of a loop in a network is detected, it is difficult to identify a coupling causing the loop. Even analysis of information indicating communication states of all of relay devices in the network can merely estimate locations of occurrence of the loop. When a continuity check is made using a ping command or the like to confirm whether the loop has occurred at locations estimated to be the locations of occurrence of the loop, processing using the ping command causes a loop and frame congestion again. Thus, it is difficult to confirm whether a loop has occurred even when the locations of occurrence of the loop are estimated. From the above, it is desirable to be able to easily identify a path in which a loop has occurred in a network.